The area of transcriptomics analysis is among the more established in computational biology, having evolved in both technology and experimental design. Transcriptomics has a strong impetus to develop sophisticated computational methods due to the large amounts of available whole-genome datasets for many species and because of powerful applications in regulatory network reconstruction as well as elucidation and modeling of cellular transcriptional responses. While gene expression microarray data can be noisy and comparisons across experiments challenging, there are a number of sophisticated methods that aid in arriving at statistically and biologically significant conclusions. As such, computational transcriptomics analysis can provide guidance for analysis of results from newer experimental technologies. More recently, search methods have been developed to identify modules of genes, which exhibit coherent expression patterns in only a subset of experimental conditions. The latest advances in these methods allow to integrate multiple data types and datasets, both experimental and computational, within a single statistical framework accounting for data confidence and relevance to specific biological questions. Such frameworks provide a unified environment for the exploration of specific biological hypothesis and for the discovery of coherent data patterns along with the evidence supporting them.